Maturation Through Interaction with Human Copper Chaperone for SOD1 (Hccs)

Human superoxide dismutase 1 (hSOD1) maturation through interaction with human copper chaperone for SOD1 (hCCS)

Lucia Bancia,b,c,1, Ivano Bertinia,b,c,1, Francesca Cantinia,b, Tatiana Kozyrevaa, Chiara Massagnia,c, Peep Palumaad, Jeffrey T. Rubinoa, and Kairit Zovod

aMagnetic Resonance Center, University of Florence, 50019 Sesto Fiorentino, Italy; bDepartment of Chemistry, University of Florence, 50019 Sesto Fiorentino, Italy; cFondazione Farmacogenomica FiorGen onlus, 50019 Sesto Fiorentino, Italy; and dDepartment of Gene Technology, Tallinn University of Technology, 12618 Tallinn, Estonia

Edited by Harry B. Gray, California Institute of Technology, Pasadena, CA, and approved July 10, 2012 (received for review May 3, 2012) Copper chaperone for superoxide dismutase 1 (SOD1), CCS, is the to be critical for CCS-SOD1 protein recognition (33–35). D2 is physiological partner for the complex mechanism of SOD1 matu- unable to bind copper (33); human D2 binds one zinc ion with the ration. We report an in vitro model for human CCS-dependent same ligands as in SOD1, which are absent in yeast D2. Zinc binding SOD1 maturation based on the study of the interactions of human is essential to human CCS (hCCS) function, most likely contributing SOD1 (hSOD1) with full-length WT human CCS (hCCS), as well as to protein stabilization (36). CCS domain 3 (D3) is a short poly- with hCCS mutants and various truncated constructs comprising peptide segment (30–40 amino acids) lacking secondary structure one or two of the protein’s three domains. The synergy between but containing a CXC motif (Cys-244 and Cys-246 in hCCS) es- electrospray ionization mass spectrometry (ESI-MS) and NMR is fully sential for CCS-dependent activation of SOD1 (29). The crystal exploited. This is an in vitro study of this process at the molecular structure of a heterodimeric complex, where an intermolecular level. Domain 1 of hCCS is necessary to load hSOD1 with Cu(I), re- disulfide bond links Cys-57 of an inactive mutant of ySOD1 and Cys- quiring the heterodimeric complex formation with hSOD1 fostered 229 (Cys-244 in hCCS) of yCCS D3, has been solved (37). It has by the interaction with domain 2. Domain 3 is responsible for the been proposed that the CCS-dependent SOD1 maturation process catalytic formation of the hSOD1 Cys-57–Cys-146 disulfide bond, involves the interaction of CCS with the reduced disulfide, zinc- which involves both hCCS Cys-244 and Cys-246 via disulfide transfer. bound form of SOD1 protein, leading to the formation of the dimeric Cu,Zn-SOD1 protein, in a process requiring oxygen (29, 38). copper transport | protein maturation | protein-protein interactions Many questions remain unanswered regarding the CCS-mediated SOD1 maturation, (e.g., it is still unknown at what point in the uperoxide dismutase 1 (SOD1) is a well-characterized process the proposed intermolecular disulfide bond is formed, if Scuproenzyme that catalyzes the dismutation of the toxic su- copper and disulfide transfer occurs separately or in a single syn- peroxide anion, a byproduct of cellular respiration, to molecular ergistic step, and what the specific roles of the various domains are). − + oxygen and hydrogen peroxide [2O2 + 2H → H2O2 + O2] (1). The goal of this work is to elucidate, at the molecular level, the CCS- In mammals, SOD1 is ubiquitously expressed in all tissues and dependent mechanism of SOD1 maturation. We have addressed primarily localized in the cytosol, although small amounts are this through NMR and electrospray ionization mass spectrometry found also in the nucleus, peroxisomes, and intermembrane space (ESI-MS) characterization of apo and copper-loaded full-length (IMS) of mitochondria (2–6). Although mature eukaryotic SOD1 WT and mutant forms of hCCS, as well as constructs of individual is a remarkably stable homodimer, the nascent monomer requires domains of hCCS and their interactions and effect on various several posttranslational modifications, including incorporation of metallated forms of hSOD1, with its disulfide bond in either an zinc and copper ions and the formation of a disulfide bond between oxidized or reduced state. We show that hSOD1 reaches its mature BIOCHEMISTRY Cys 57 and Cys 146, before reaching the active dimeric form (7–10). copper-loaded form in vitro only when the disulfide-reduced zinc- Despite having a wealth of structural information regarding apo bound form of hSOD1 is treated with Cu(I)-bound full-length and various metallated forms of SOD1 (11–13), the mechanisms of hCCS, whereas truncated forms of hCCS are not able to produce posttranslational modification are still obscure. Understanding these a mature dimeric Cu,Zn-hSOD1 protein. The knowledge acquired mechanisms is of critical importance because immature forms of using a combination of these complementary techniques (NMR and SOD1 are believed to be linked to the neurodegenerative disease ESI-MS) allowed us to reinterpret in a comprehensive manner the – familial amyotrophic lateral sclerosis (14 17). Although chaperone- complete pathway of SOD1 maturation by CCS. independent maturation of SOD1 has been reported in humans (hSOD1) (18, 19) and other organisms (20–23), the majority of the Results SOD1 activation within the cell, or 100% of activation in the case of In vitro, hSOD1 reaches its mature, copper-loaded form when the † yeast SOD1 (ySOD1), is via the SOD1-specific chaperone protein, reduced, zinc-containing form of hSOD1 (E*,Zn-hSOD1SH )is copper chaperone for SOD1 (CCS) (24–28). CCS also likely exhibits disulfide isomerase activity (29, 30). However, the molecular mechanism of CCS-dependent maturation of SOD1 is still unknown. Author contributions: L.B., I.B., and P.P. designed research; F.C., T.K., C.M., J.T.R., and K.Z. CCS is almost ubiquitously expressed in eukaryotes, with com- performed research; F.C., T.K., J.T.R., and K.Z. analyzed data; and L.B., I.B., F.C., T.K., P.P., parable cellular distribution but lower expression levels than SOD1 J.T.R., and K.Z. wrote the paper. (27, 31). CCS is a dimeric protein, with each monomer constituted The authors declare no conflict of interest. by three distinct domains. The human CCS N-terminal domain This article is a PNAS Direct Submission. 1 (D1) is homologous and shows the same fold of the Atx1-like 1To whom correspondence may be addressed. E-mail: [email protected]fi.it or ivanobertini@ metallochaperones and harbors the MXCXXC copper-binding cerm.unifi.it. motif typical of this family of proteins. Human D1 is essential for This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. hSOD1 copper acquisition in vivo (32), whereas yeast D1 appears to 1073/pnas.1207493109/-/DCSupplemental. be necessary only under copper-limiting conditions (33). CCS do- *E indicates the copper-binding site is empty. † main 2 (D2) is structurally similar to SOD1 and has been postulated A superscripted SH indicates Cys-57 and Cys-146 are both reduced.

www.pnas.org/cgi/doi/10.1073/pnas.1207493109 PNAS | August 21, 2012 | vol. 109 | no. 34 | 13555–13560 Downloaded by guest on September 29, 2021 detected by ESI-MS (Fig. S4). Upon titration of E,Zn-hSOD1SH with Cu(I)-hCCSD2,3, a substantial amount of hCCSD2,3 was found in complex with E,Zn-hSODSH via ESI-MS but no copper transfer was observed (Fig. 3), indicating that the copper weakly bound to D3 cannot be transferred to SOD1, despite their interaction. Titration of E,Zn-hSOD1SH with hCCSD2 also resulted in the formation of the heterodimeric complex. Therefore, in any CCS/ SOD1 mixture in which D2 was present, interaction between monomeric forms of hSOD1 andhCCS, resulting in the formationof 1:1 heterodimeric complexes, was detected by ESI-MS. Disulfide bond formation in hSOD1 and dimerization of protein were not observed for any of these titrations involving single- or double-domain hCCS constructs (Fig. 2C). No copper transfer was observed by NMR when reduced apo hSOD1 (E,E-hSOD1SH), lacking both copper and zinc, was incubated with Cu(I)-hCCSD1,2 (Fig. S5A) even though formation of the heterodimer has been detected via ESI-MS (Fig. S5B). Overall, these data indicate that all three domains of hCCS are required for obtaining the mature hSOD1: copper transfer from hCCS to hSOD1 is occurring via D1, whereas D2 has a critical role in protein-protein recognition, which is a necessary interaction for D1 to transfer copper to SOD1. Also, zinc binding to hSOD1 before fi Fig. 1. Incubation of E,Zn-hSOD1SH with full-length Cu(I)-hCCS resulted in copper binding could be a necessary step to induce a de ned and mature dimeric Cu,Zn-hSOD1S-S. ESI-MS spectra of E,Zn-hSOD1SH (10 μM) optimal conformation for copper coordination. incubated for 2 min (A) and for 1 h (B) with full-length Cu(I)-hCCS (10 μM) are The overall behavior is different when various hCCS constructs S-S shown. Conditions are as follows: 20 mM ammonium acetate, 0.5 mM DTT, were titrated with oxidized forms of hSOD1 . Copper-loaded pH 7.5, at 25 °C. The +8 charge state is presented for E,Zn-hSOD1SH mono- hCCS single- or double-domain constructs (hCCSD1, hCCSD1,2, mer, and the +11 charge state is presented for dimer. The metal content of or hCCSD2,3) were either inefficient or totally unable to metallate the protein is also indicated. the oxidized form of E,Zn-hSOD1S-S (Fig. S6). The heterodimeric complex between oxidized E,Zn-hSOD1S-S and hCCSD2,3 was present at a much lower extent compared with incubation with the treated with full-length Cu(I)-hCCS, resulting in both copper reduced form (Fig. S6 A and D), indicating that these complexes transfer and SOD1 disulfide bond formation. Indeed, incubation of fi SH are highly populated only when the disul de bond is reduced, as E,Zn-hSOD1 , which is monomeric, with Cu(I)-hCCS results in S-S S-S‡ expected. Indeed, E,Zn-hSOD1 is predominantly present in the the formation of the mature, oxidized dimeric Cu,Zn-hSOD1 dimeric form; thus, dissociation of the homodimer into monomers fi within 1 h (Fig. 1). Copper transfer was less ef cient using the oxi- able to interact with hCCS may be disfavored. The inability to dized, zinc-bound form of hSOD1 (E,Zn-hSOD1S-S), which is di- S-S metallate the oxidized form of hSOD1 reinforces the hypothesis meric, reaching the fully metallated form (Cu,Zn-hSOD1 ), that the posttranslational maturation of hSOD1 is a stepwise together with other partially metallated forms, only after incubation process, requiring metallation before disulfide formation. The for around 13 h (Fig. S1). E,Zn-hSOD1SH also binds Cu(I) when fi SH disul de bond has a fundamental role in stabilizing the residues in provided as a free ion in solution, leading to Cu,Zn-hSOD1 after the copper-binding region. The hSOD1 disulfide may prevent fi 1 h of incubation. However, no disul de bond was formed, leaving copper loading via the CCS-dependent mechanism, because for- the protein in the reduced monomeric state (Fig. S2). mation of the disulfide bond reduces the accessibility of the metal- Full-length hCCS has previously been reported in a dimeric binding catalytic site (42). Alternatively, hSOD1 Cys-57 may me- state (39, 40). Each monomer can bind Cu(I) at two different diate copper transfer; thus, oxidization may inhibit copper loading. fi sites with quite different af nity. Cu(I) binds to the MXCXXC To dissect the role of the various domains of CCS further, E, SH copper-binding motif of D1, a typical Cu(I) chaperone, with a Kd Zn-hSOD1 was titrated with two hCCS constructs either step- −18 of 2.34 ± 0.02 × 10 M, as determined here by ESI-MS using wise or simultaneously. E,Zn-hSOD1SH was titrated with Cu an hCCS construct in which only D1 is expressed (hCCSD1). (I)-hCCSD1,2, thus obtaining Cu,Zn-hSOD1SH, and was sub- This value is in agreement with recently published data (40). Cu sequently exposed to air. Even in the presence of oxygen, no for- (I) can also coordinate to the CXC motif of the unstructured D3 mation of the disulfide bond was observed until addition of (41) but with a much lower affinity with respect to D1 (40). hCCSD2,3 to the reaction mixture, albeit oxidation was only par- Actually, no binding to an hCCS construct containing D2 and D3 tial. Similar results were also observed in the absence of air when (hCCSD2,3) is observed in the presence of 0.5 mM DTT. E,Zn-hSOD1SH, titrated with Cu(I)-hCCSD1,2, was subsequently To dissect the mechanism of hSOD1 maturation by hCCS, par- titrated with oxidized hCCSD2,3 (i.e., the CXC motif of D3 was ticularly therole of the three domains during the maturationprocess, oxidized). These data indicates that D3 is needed in disulfide bond we first analyzed the interactions of various hCCS constructs com- formation. The role of hCCS D3 in disulfide transfer, as well as D1 posed of single (hCCSD1 and hCCSD2) or double (hCCSD1,2 and in copper transfer, was also confirmed by titrations of E, hCCSD2,3) domains in apo and/or metallated forms, with both ox- Zn-hSOD1SH with mixtures of Cu(I)-hCCSD1 and oxidized idized and reduced hSOD1. Neither complex formation nor copper hCCSD2,3 or with mixtures of hCCSD1 and Cu(I)-hCCSD2,3. transfer was detected (by ESI-MS) when Cu(I)-hCCSD1 was added Copper transfer and formation of the disulfide bond were ob- to E,Zn-hSOD1SH (Fig. S3), and only partial copper transfer was served only in the former case [i.e., in the presence of both Cu(I)- observed (by NMR) after prolonged incubation (Fig. 2). When E, hCCSD1 and hCCSD2,3], although the process was not complete. Zn-hSOD1SH was incubated with Cu(I)-hCCSD1,2, on the contrary, To confirm our results regarding the role of D1 in Cu(I) complete copper transfer was observed by NMR (Fig. 2), whereas transfer and D3 in disulfide formation, NMR was used to monitor both copper transfer and formation of a heterodimer complex were E,Zn-hSOD1SH titrated with full-length Cu(I)-hCCS in which the cysteines of the D3 CXC motif were replaced with Ala either in single C244A and C246A mutations or in a double C244/246A ‡ A superscripted S-S indicates the presence of the C57-C146 disulfide bond. mutation, initially monitored anaerobically. For all three mutants,

13556 | www.pnas.org/cgi/doi/10.1073/pnas.1207493109 Banci et al. Downloaded by guest on September 29, 2021 Fig. 2. Transfer of Cu(I) to the hSOD1 is complete in the presence of Cu(I)-hCCSD1,2 and incomplete in the presence of Cu(I)-hCCSD1, with both proteins unable to promote the formation of the disulﬁde bond of hSOD1. (A) 1H-15N heteronuclear single quantum correlation spectroscopy (HSQC) spectrum of E,Zn-hSOD1SH.(B and C) Overlay of selected regions of the 1H-15N HSQC spectra of E,Zn- hSOD1SH (black), E,Zn-hSOD1SH in the presence of Cu(I)-hCCSD1,2 (red), and E,Zn-hSOD1SH in the presence of Cu(I)-hCCSD1 (blue). The chemical shift assignments of E,Zn-hSOD1SH and Cu,Zn-hSOD1S-S are available (42, 49). In B, the amide signal of residue Gly-73 is demonstrative of copper binding because it is located in proximity to the catalytic metal-binding site, whereas in C, the chemical shifts of side chain signals of Asn-53 are affected by both copper binding and oxidation. Asn-53 is located in loop 4 of hSOD1 and faces the C57-C146 disulﬁde on its formation.

Cu(I) transfer was observed but no disulfide formation occurred, the complex with SOD1 required to facilitate copper transfer. with the latter only being partially formed on prolonged in- This is a different behavior with respect to copper proteins that cubation (15 h) with air after copper transfer was complete. This receive the metal from Atx1-like chaperones, which show the same is different with respect to what has been observed by NMR in fold and Cu(I)-binding site of hCCSD1. However, the partners of the case of WT full-length Cu(I)-hCCS in a similar experiment, the Atx1-like chaperones are completely different from SOD1, not with almost complete oxidation (80%) observed within the 3.75-h only in terms of overall fold but in terms of the metal coordination time frame of the experiment (Fig. 4). Therefore, it appears that sites and their solvent accessibility (43–46). hCCSD2, in addition only full-length Cu(I)-hCCS, with the concerted involvement of to possessing the same fold of SOD1, contains a hydrophobic all three domains, is able to produce mature hSOD1 effectively, region of the β-barrel that is perfectly suited to interact with the starting only from the reduced, zinc-bound form of SOD1. hydrophobic region of the β-barrel of the SOD1 monomer. This The midpoint redox potential of the cysteine pairs in hCCSD1 is − interaction would then bring hCCSD1 into an optimal orientation 243.0 ± 0.5 mV, whereas it is −270.0 ± 1.7 mV for hCCSD2,3 (Fig. for copper transfer. Finally, oxidation of the hSOD1 intra- S7A). The determination of the midpoint redox potential of E, molecular disulfide bond occurs only in the presence of D3. Zn-hSOD1 and Cu,Zn-hSOD1 was complicated by the occurrence of Furthermore, using NMR to compare the maturation of hSOD1 monomer/dimer equilibrium between the reduced and oxidized utilizing both WT full-length hCCS and Cys-to-Ala mutants of D3, species, respectively. From the dependence of intensities of the mo- we have shown that when the CXC motif is compromised, Cu(I) nomeric peak in ESI-MS spectra, and taking into account the redox transfer is maintained, whereas disulfide formation is negated. This potential of the solution, a value of −351.6 ± 1.5 mV was determined confirms that D1 is responsible for Cu(I) transfer during hCCS-

for E,Zn-hSOD1 and a value of −337 ± 1.7 mV was determined for dependent hSOD1 maturation. These results also indicate that both BIOCHEMISTRY Cu,Zn-hSOD1 (Fig. S7 B and C). These reduction potentials are cysteines in the D3 CXC motif are required for disulfide isomerase consistent with the role of hCCS as an oxidant of hSOD1. activity, because both the single C244A and C246A mutants reduce the rate of Cu,Zn-hSOD1SH oxidation similar to the double-C244/ Discussion 246A mutant (Fig. 4B). This claim is further supported by titrations The present work led us to propose an updated model of hCCS- of Cu,Zn-hSOD1SH with oxidized hCCSD2,3, where the disulfide in dependent hSOD1 maturation involving at least six steps: zinc the CXC motif is transferred to hSOD1 presumably via the for- acquisition from an unknown source; hCCS-hSOD1 heterodimer mation of an intramolecular disulfide bond between hCCS C244 formation; hCCS-to-hSOD1 copper transfer via D1; transfer of the and hSOD1 C57 as previously reported in the yeast system (29). We hCCS C244-C246 disulfide bond to hSOD1 C57-C146, presumably determined the reduction potential for E,Zn-hSOD1 and Cu,Zn- via formation of an intermolecular disulfide between C244 of hCCS hSOD1 to be −351.6 ± 1.5 mV and −337 ± 1.7 mV, respectively. and C57 of hSOD1; and, finally, dimerization of the mature hSOD1 Although the reduction potential of the cellular compartments monomer to the active form of the protein (Fig. 5). where hSOD1 is found, cytoplasm and the mitochondrial IMS, Using NMR and ESI-MS methods, and exploiting a number of should be able to oxidize the disulfide bond (47), we have observed constructs that feature different combinations of hCCS domains, that in the absence of hCCS (Fig. 4), the oxidation process in vitro is we have been able to show that only hCCS constructs containing kinetically slow, as previously reported for yeast (29). This suggests copper-loaded D1 (hCCSD1,2 and, to a lesser extent, hCCSD1) that SOD1 disulfide formation is under kinetic control. The specific are able to transfer copper to hSOD1, whereas a construct protein-protein interaction between one subunit of hCCS and one containing copper-loaded D3 (hCCSD2,3) was not. No copper subunit of hSOD1 might be crucial not only for copper(I) transfer transfer from constructs containing copper-loaded D1 to con- but for the disulfide exchange reaction. Protein-protein interactions structs containing D3 was observed. This observation is fully could position Cys57 of the disulfide-reduced hSOD1, located in consistent with the much higher Cu(I) affinity of D1 with respect a solvent-exposed loop at the heterodimer interface, in the correct to that of D3, which thermodynamically disfavors Cu(I) transfers orientation to promote the formation of an intermolecular disulfide from D1 to D3 (40). The metal transfer is efficient and complete bond with CCS D3, which may then rapidly evolve in the formation only in the presence of D2, because D1 alone is unable to form of the intramolecular bond.

Banci et al. PNAS | August 21, 2012 | vol. 109 | no. 34 | 13557 Downloaded by guest on September 29, 2021 Although this supports our ﬁnding that D1 is responsible for copper transfer, this study also reports that hSOD1 activity and copper in- − − corporation were negated when the hCCS / cells were transfected with a full-length hCCS mutant containing a mutation in the D3 CXC-binding motif. This supported the proposal that D3 is required for the insertion of copper transfer into hSOD1, at variance with what we have observed in the present work. Similar studies performed on a yeast system lacking CCS have shown a complete loss of ySOD1 activity in the absence of D3, which suggested that D3 was absolutely essential for CCS maturation of SOD1. The yeast study also reported that D1 was only essential under copper-limiting conditions, and separate experiments in humans reported that D1 mutations only result in decreased hSOD1 activity (48). Although the literature appears consistent in regard to the essential role of D3

Fig. 3. No Cu(I) transfer to the hSOD1 is observed in the presence of Cu(I)- hCCSD2,3. ESI-MS spectra of E,Zn-hSOD1SH (10 μM) incubated for 2 min (A) and for 1 h (B) with Cu(I)-hCCSD2,3 (10 μM) are shown. Conditions are as follows: 20 mM ammonium acetate, 0.5 mM DTT, pH 7.5, at 25 °C. For E, Zn-hSOD1SH the +9, +8, and +7 monomer charge states are present from left to right, respectively; Peaks associated with hCCSD2,3 are indicated; the +13 and +12 charge states for heterodimer complexes are labeled ac- cordingly. The metal contents are also indicated. (C) Overlay of selected regions of the 1H-15N heteronuclear single quantum correlation spectroscopy spectra of E,Zn-hSOD1SH (black), E,Zn-hSOD1SH in the presence of Cu(I)- hCCSD1,2 (red), and E,Zn-hSOD1SH in the presence of Cu(I)-hCCSD2,3 (blue).

Efficient formation of the mature hSOD1 was only obtained from incubation with full-length Cu(I)-hCCS. Although copper transfer was efficient on incubation of E,Zn-hSOD1SH with Cu(I)- hCCSD1,2, only partial oxidation was observed with subsequent addition of oxidized hCCSD2,3. In this fashion, the copper transfer and disulfide transfer are occurring in two separate steps, with each requiring its own separate D2-dependent recognition event to form Fig. 4. Both cysteines in the D3 CXC motif are required for disulfide isom- the heterodimer. This proposed mechanism would be inherently erase activity. (A) Overlay of selected regions of the 1H-15N HSQC spectra less efficient compared with the mechanism using the full-length of E,Zn-hSOD1SH (black), E,Zn-hSOD1SH in the presence of full-length Cu(I)- protein, where both processes occur in a coordinated way, requiring C244/246A mutant (dark gray), and E,Zn-hSOD1SH in the presence of WT full- only one binding event. This may be why nature has evolved CCS length Cu(I)-hCCS (light gray) after 225 min. (B) Percentage of reduced with three domains able to carry out both processes in a synergistic hSOD1 remaining after opened to air for 10 min and spectra collected for SH fashion, as opposed to relying on two proteins to accomplish the 225 min. E,Zn-hSOD1 is shown alone, after initial anaerobic incubation + separate copper and disulfide transfer functions. with WT Cu(I)-hCCS ( WT) and copper-loaded forms of the C244/246A mutation (+C244/246A), C244A mutation (+C244A), and C246A mutation (+ In a previous study examining the roles of the individual domains SH C246A). In each case, where EZn-hSOD1 was incubated with copper-loaded of hCCS at the cellular level, neither SOD1 activity nor copper −/− CCS (both WT and mutant forms), 100% copper transfer was observed be- incorporation into SOD1 was observed in hCCS mammalian fore opening the samples to air. For each experiment, the intensity of three cells transfected with hCCSD2,3 or with a full-length hCCS mutant peaks associated with the reduced form of the protein was averaged and with Cys-to-Ser mutations in the D1 MXCXXC motif (32). normalized to the corresponding peak in the absence of oxidation.

13558 | www.pnas.org/cgi/doi/10.1073/pnas.1207493109 Banci et al. Downloaded by guest on September 29, 2021 Fig. 5. Schematic mechanism of CCS-dependent hSOD1 maturation in vitro. (1) Nascent hSOD1 (E,E-hSODSH) acquires zinc from an unknown source, pro- ducing E,Zn-hSOD1SH. (2) The hSOD1SH monomer forms a heterodimer with a monomer Cu(I)-hCCS. The current study suggests that copper transfer (3) occurs before formation of the intermolecular (4), and then hSOD1 intramolecular (5), disulﬁde bond, resulting in the mature monomer (6) that dimerizes to the active state. Although oxygen in vitro is required in the process before step 4, it is currently unknown precisely at which step this occurs. [Note: E,Zn-hSOD1SH can obtain Cu(I) in a CCS-independent manner.]

for SOD1 maturation, the necessity for and role of D1 remained This work is a study of the maturation of hSOD1 via hCCS at the inconclusive (30, 32, 33, 41, 48). The previous reported findings were molecular level. Using a combination of NMR and ESI-MS, we have principally based on assays measuring for overall SOD1 activity elucidated various steps in the process of CCS-dependent matu- (which requires complete SOD1 maturation); thus, they were unable ration of hSOD1, specifically the roles of the individual domains to monitor separately for copper acquisition and disulfide bond during copper acquisition and disulfide formation processes. formation, as accomplished in this work using NMR. Based on our findings, and consistent with what has been Materials and Methods reported in the literature, we can propose that the SOD1 de- hSOD1, Full-Length hCCS, and hCCS Constructs. Expression of full-length hCCS pendence on D3 for maturation is the result of its role in disulfide and constructs, as well hSOD1, including metallation techniques, is described isomerase activity and not a copper transport function, whereas in SI Materials and Methods. D1 is responsible for Cu(I) insertion. These results are particularly interesting considering that some organisms lack the D1 ESI-MS Measurements for Protein-Protein Interactions and Copper-Binding Affinity. For all protein samples, including hSOD1, full-length CCS, and

MXCXXC copper-binding motif, including Drosophila mela- BIOCHEMISTRY fi constructs, protein-protein interactions were monitored after various time nogaster and Anopheles gambiae (23) and the ssion yeast Sac- points from 2 min to 13 h, depending on the experiment. For copper de- charomyces pombe (21). Because CCS from D. melanogaster is able termination of copper-binding affinity, increasing concentrations of DTT to activate D. melanogaster SOD1 and ySOD1 fully without were added to 10-μM samples of the apo or Cu(I)-bound form of hCCSD1 and exhibiting copper chaperone activity (23), it appears that the pri- hCCSD2,3 constructs. Samples were then incubated for 2 min and finally mary function of CCS is that of a disulfide isomerase. This then analyzed by ESI-MS. Further experimental details and instrument settings poses the question of how SOD1 is obtaining copper in the ab- are available in SI Materials and Methods. sence of a functional D1. We speculate that the most likely answer is via the same mechanism as in the CCS-independent maturation Determination of Midpoint Redox Potential of Various CCS Domains, E,Zn- of SOD1, which is currently attributed to a copper–glutathione hSOD1, and Cu,Zn-hSOD1. For determination of the midpoint redox potential for hCCSD1 and hCCSD23, the fully reduced hCCS domains were incubated in complex as observed in vitro in yeast (27). Also, human CCS-in- β-mercaptoethanol–based (BME) redox buffers at room temperature for 1 h, dependent SOD1 maturation is known to require glutathione (19). which was sufficient to reach the redox equilibrium between the oxidized Thus, there are at least three mechanisms for SOD1 maturation: and reduced protein forms. The BME buffer with a total concentration of a first mechanism whereby SOD1 is dependent on CCS for both 5 mM BME was used. For determination of the midpoint redox potential for copper transfer (via D1) and disulfide transfer (via D3), such as is E,Zn-hSOD1 and Cu,Zn-hSOD1, the fully oxidized proteins were incubated in observed for yeast under copper-limiting conditions (33); a second DTT-based redox buffers at room temperature for 1 h with a total concen- mechanism whereby SOD1 acquires copper from another source tration of 5 mM DTT. Further experimental details, instrument settings, and and then the disulfide from CCS (via D3), such as is likely for D. mathematical determination of midpoint redox potentials using the Nernst melanogaster (23); and a third mechanism comprising a completely equation are given in SI Materials and Methods. CCS-independent system, such as is observed for Caenorhabditis NMR Experiments. The oxidized or reduced form of E,E-hSOD1 and E,Zn- elegans,whichlacksCCSaltogether(24),aswellasinorganisms hSOD1bwas titrated with unlabeled full-length hCCS or constructs and moni- in which the chaperone-dependent systems are also present tored by 1H-15N heteronuclear single quantum correlation spectroscopy NMR. (19). For organisms in which both CCS-dependent and independent The chemical shift assignments of E,Zn-hSOD1SH and Cu,Zn-hSOD1S-S are mechanisms have been described, such as for humans, CCS- available (42, 49). Details regarding instrumentation and data processing are dependent maturation is more efficient (22). given in SI Materials and Methods.

Banci et al. PNAS | August 21, 2012 | vol. 109 | no. 34 | 13559 Downloaded by guest on September 29, 2021 ACKNOWLEDGMENTS. This study was supported by the Programmi di NMR (European Commission Framework Programme 7 e-Infrastructure Ricerca di Rilevante Interesse Nazionale (PRIN) (2009FAKHZT_001 “Biologia grant, Contract No. 261863, Estonian Science Foundation Grant No. strutturale meccanicistica: avanzamenti metodologici e biologici”), Bio- 8811, and Ente Cassa di Risparmio di Firenze.

1. McCord JM, Fridovich I (1969) Superoxide dismutase. An enzymic function for er- 26. Culotta VC, et al. (1997) The copper chaperone for superoxide dismutase. J Biol Chem ythrocuprein (hemocuprein). J Biol Chem 244:6049–6055. 272:23469–23472. 2. Kobayashi T, et al. (1993) Ultrastructural localization of superoxide dismutase in hu- 27. Rae TD, Schmidt PJ, Pufahl RA, Culotta VC, O’Halloran TV (1999) Undetectable in- man skin. Acta Derm Venereol 73(1):41–45. tracellular free copper: The requirement of a copper chaperone for superoxide dis- 3. Crapo JD, Oury T, Rabouille C, Slot JW, Chang LY (1992) Copper,zinc superoxide dismutase. Science 284:805–808. mutase is primarily a cytosolic protein in human cells. Proc Natl Acad Sci USA 89: 28. Rae TD, Torres AS, Pufahl RA, O’Halloran TV (2001) Mechanism of Cu,Zn-superoxide 10405–10409. dismutase activation by the human metallochaperone hCCS. J Biol Chem 276: 4. Lindenau J, Noack H, Possel H, Asayama K, Wolf G (2000) Cellular distribution of 5166–5176. – superoxide dismutases in the rat CNS. Glia 29(1):25 34. 29. Furukawa Y, Torres AS, O’Halloran TV (2004) Oxygen-induced maturation of SOD1: A ’ 5. Field LS, Furukawa Y, O Halloran TV, Culotta VC (2003) Factors controlling the uptake key role for disulfide formation by the copper chaperone CCS. EMBO J 23:2872–2881. of yeast copper/zinc superoxide dismutase into mitochondria. J Biol Chem 278: 30. Proescher JB, Son M, Elliott JL, Culotta VC (2008) Biological effects of CCS in the ab- – 28052 28059. sence of SOD1 enzyme activation: Implications for disease in a mouse model for ALS. 6. Sturtz LA, Diekert K, Jensen LT, Lill R, Culotta VC (2001) A fraction of yeast Cu,Zn- Hum Mol Genet 17:1728–1737. superoxide dismutase and its metallochaperone, CCS, localize to the intermembrane 31. Rothstein JD, et al. (1999) The copper chaperone CCS is abundant in neurons and space of mitochondria. A physiological role for SOD1 in guarding against mitochon- astrocytes in human and rodent brain. J Neurochem 72:422–429. drial oxidative damage. J Biol Chem 276:38084–38089. 32. Caruano-Yzermans AL, Bartnikas TB, Gitlin JD (2006) Mechanisms of the copper-de- 7. Forman HJ, Fridovich I (1973) On the stability of bovine superoxide dismutase. The pendent turnover of the copper chaperone for superoxide dismutase. J Biol Chem effects of metals. J Biol Chem 248:2645–2649. 281:13581–13587. 8. Roe JA, et al. (1988) Differential scanning calorimetry of Cu,Zn-superoxide dismutase, 33. Schmidt PJ, et al. (1999) Multiple protein domains contribute to the action of the the apoprotein, and its zinc-substituted derivatives. Biochemistry 27:950–958. copper chaperone for superoxide dismutase. J Biol Chem 274:23719–23725. 9. Bertini I, Mangani S, Viezzoli MS (1998) Structure and Properties of Copper-Zinc 34. Lamb AL, et al. (1999) Crystal structure of the copper chaperone for superoxide dis- Superoxide Dismutases. Advances in Inorganic Chemistry (Academic, San Diego), Vol – 45, pp 127–250. mutase. Nat Struct Biol 6:724 729. ’ 10. Arnesano F, et al. (2004) The unusually stable quaternary structure of human Cu,Zn- 35. Lamb AL, Wernimont AK, Pufahl RA, O Halloran TV, Rosenzweig AC (2000) Crystal superoxide dismutase 1 is controlled by both metal occupancy and disulfide status. structure of the second domain of the human copper chaperone for superoxide – J Biol Chem 279:47998–48003. dismutase. Biochemistry 39:1589 1595. 11. Banci L, et al. (1998) Solution structure of reduced monomeric Q133M2 copper, zinc 36. Endo T, Fujii T, Sato K, Taniguchi N, Fujii J (2000) A pivotal role of Zn-binding residues superoxide dismutase (SOD). Why is SOD a dimeric enzyme? Biochemistry 37: in the function of the copper chaperone for SOD1. Biochem Biophys Res Commun 11780–11791. 276:999–1004. 12. Banci L, Bertini I, Cantini F, D’Onofrio M, Viezzoli MS (2002) Structure and dynamics of 37. Lamb AL, Torres AS, O’Halloran TV, Rosenzweig AC (2001) Heterodimeric structure of copper-free SOD: The protein before binding copper. Protein Sci 11:2479–2492. superoxide dismutase in complex with its metallochaperone. Nat Struct Biol 8: 13. Banci L, Bertini I, Cramaro F, Del Conte R, Viezzoli MS (2003) Solution structure of Apo 751–755. Cu,Zn superoxide dismutase: Role of metal ions in protein folding. Biochemistry 38. Banci L, Bertini I, Cantini F, Ciofi-Baffoni S (2010) Cellular copper distribution: A 42:9543–9553. mechanistic systems biology approach. Cell Mol Life Sci 67:2563–2589. 14. Wang J, et al. (2003) Copper-binding-site-null SOD1 causes ALS in transgenic mice: 39. Wright GSA, Hasnain SS, Grossmann JG (2011) The structural plasticity of the human Aggregates of non-native SOD1 delineate a common feature. Hum Mol Genet copper chaperone for SOD1: Insights from combined size-exclusion chromatographic 12:2753–2764. and solution X-ray scattering studies. Biochem J 439(1):39–44. 15. Banci L, et al. (2007) Metal-free superoxide dismutase forms soluble oligomers under 40. Allen S, Badarau A, Dennison C (2012) Cu(I) affinities of the domain 1 and 3 sites in physiological conditions: A possible general mechanism for familial ALS. Proc Natl the human metallochaperone for Cu,Zn-superoxide dismutase. Biochemistry 51: Acad Sci USA 104:11263–11267. 1439–1448. 16. Lelie HL, et al. (2011) Copper and zinc metallation status of copper-zinc superoxide 41. Stasser JP, Eisses JF, Barry AN, Kaplan JH, Blackburn NJ (2005) Cysteine-to-serine dismutase from amyotrophic lateral sclerosis transgenic mice. J Biol Chem 286: mutants of the human copper chaperone for superoxide dismutase reveal a copper – 2795 2806. cluster at a domain III dimer interface. Biochemistry 44:3143–3152. “fi ” 17. Banci L, et al. (2010) NMR characterization of a bril-ready state of demetalated 42. Banci L, Bertini I, Cantini F, D’Amelio N, Gaggelli E (2006) Human SOD1 before har- – wild-type superoxide dismutase. J Am Chem Soc 133:345 349. boring the catalytic metal: Solution structure of copper-depleted, disulfide-reduced 18. Corson LB, Strain JJ, Culotta VC, Cleveland DW (1998) Chaperone-facilitated copper form. J Biol Chem 281:2333–2337. binding is a property common to several classes of familial amyotrophic lateral scle- 43. Banci L, et al. (2005) A NMR study of the interaction of a three-domain construct of rosis-linked superoxide dismutase mutants. Proc Natl Acad Sci USA 95:6361–6366. ATP7A with copper(I) and copper(I)-HAH1: The interplay of domains. J Biol Chem 280: 19. Carroll MC, et al. (2004) Mechanisms for activating Cu- and Zn-containing superoxide 38259–38263. dismutase in the absence of the CCS Cu chaperone. Proc Natl Acad Sci USA 101: 44. Banci L, et al. (2006) The Atx1-Ccc2 complex is a metal-mediated protein-protein in- 5964–5969. teraction. Nat Chem Biol 2:367–368. 20. Wong PC, et al. (2000) Copper chaperone for superoxide dismutase is essential to 45. Achila D, et al. (2006) Structure of human Wilson protein domains 5 and 6 and their activate mammalian Cu/Zn superoxide dismutase. Proc Natl Acad Sci USA 97: interplay with domain 4 and the copper chaperone HAH1 in copper uptake. Proc Natl 2886–2891. – 21. Laliberté J, et al. (2004) The Schizosaccharomyces pombe Pccs protein functions in Acad Sci USA 103:5729 5734. both copper trafficking and metal detoxification pathways. J Biol Chem 279: 46. Banci L, et al. (2009) An NMR study of the interaction of the N-terminal cytoplasmic – 28744–28755. tail of the Wilson disease protein with copper(I)-HAH1. J Biol Chem 284:9354 9360. fi 22. Jensen LT, Culotta VC (2005) Activation of CuZn superoxide dismutases from 47. Kemp M, Go YM, Jones DP (2008) Nonequilibrium thermodynamics of thiol/disul de Caenorhabditis elegans does not require the copper chaperone CCS. J Biol Chem 280: redox systems: A perspective on redox systems biology. Free Radic Biol Med 44: 41373–41379. 921–937. 23. Kirby K, et al. (2008) Instability of superoxide dismutase 1 of Drosophila in mutants 48. Stasser JP, Siluvai GS, Barry AN, Blackburn NJ (2007) A multinuclear copper(I) cluster deficient for its cognate copper chaperone. J Biol Chem 283:35393–35401. forms the dimerization interface in copper-loaded human copper chaperone for su- 24. Leitch JM, et al. (2009) Activation of Cu,Zn-superoxide dismutase in the absence of peroxide dismutase. Biochemistry 46:11845–11856. oxygen and the copper chaperone CCS. J Biol Chem 284:21863–21871. 49. Banci L, Bertini I, Cramaro F, Del Conte R, Viezzoli MS (2002) The solution structure of 25. Leitch JM, Yick PJ, Culotta VC (2009) The right to choose: Multiple pathways for reduced dimeric copper zinc superoxide dismutase. The structural effects of di- activating copper,zinc superoxide dismutase. J Biol Chem 284:24679–24683. merization. Eur J Biochem 269:1905–1915.

13560 | www.pnas.org/cgi/doi/10.1073/pnas.1207493109 Banci et al. Downloaded by guest on September 29, 2021